Reliability Block Diagrams to Model Disease Management

Background and objectives. Studies of diagnostic or therapeutic procedures in the management of any given disease tend to focus on one particular aspect of the disease and ignore the interaction between the multitude of factors that determine its final outcome. The present article introduces a mathematical model that accounts for the joint contribution of various medical and non-medical components to the overall dis ease outcome. Methods. A reliability block diagram is used to model patient compli ance, endoscopic screening, and surgical therapy for dysplasia in Barrett's esophagus. Results. The overall probability of a patient with a Barrett's esophagus to comply with a screening program, be correctly diagnosed with dysplasia, and undergo successful therapy is 37%. The reduction in the overall success rate, despite the fact that the majority of components are assumed to function with reliability rates of 80% or more, is a reflection of the multitude of serial subsystems involved in disease management. Each serial component influences the overall success rate in a linear fashion. Building multiple parallel pathways into the screening program raises its overall success rate to 91%. Parallel arrangements render systems less sensitive to diagnostic or thera peutic failures. Conclusions. A reliability block diagram provides the means to model the contributions of many heterogeneous factors to disease outcome. Since no medical system functions perfectly, redundancy provided by parallel subsystems assures a greater overall reliability. Key words: Barrett's esophagus; decision analysis; outcome research; reliability block diagram; sensitivity analysis; screening endoscopy. (Med De cis Making 1999;19:180-185)